Notes from a talk on use of colour in science and how current methods are fairly inaccessible.
NB. Saying colour-blind is somewhat derogatory (also generally incorrect because achromatopsia is super-rare), so the better way is colour-vision impaired or, perhaps, spectrally-challenged.
Colour vision impairment simulators: www.vischeck.com http://colororacle.org
This site seems to have lots of different information and tools, but is also extremely chaotic: mkweb.bcgsc.ca/colorblind
In R:
- RColorBrewer has some option called
display.brewer.all(colorblindFriendly = TRUE)
- the
viridis
package has several colour palette options (magma, plasma, inferno) colorspace::terrain_hcl()
((for some reason my notes say I don't like this, so might not look very good))dichromat
package
In Python:
brewer2mpl
matplotlib
Github for colormap, apparently the people behind viridis
and matplotlib
.
General tips from the talk:
- Blue is good, sky blue is good for highlighting
- Use magenta, vermillion, or orange instead of red
- Never use red on black
- Avoid green -> use blue-green instead
Avoiding Color - Nature Methods paper
Guidelines for preparing color figures for everyone including the colorblind - Pharmacological Research paper